Sensitivity and specificity of a procedure for early human screening of novel smoking cessation medications

Abstract

Background and aim: It is important to find economical methods in early Phase 2 studies to screen drugs potentially useful to aid smoking cessation. A method has been developed that detects efficacy of varenicline and nicotine patch. This study aimed to evaluate whether the method would detect the efficacy of bupropion and identify correctly the lack of efficacy of modafinil.

Design: Using a within-subject double cross-over design, smokers attempted to quit during each treatment, with bupropion (150 mg b.i.d.), modafinil [100 mg twice daily (b.i.d.)] or placebo (double-blind, counterbalanced order). In each of three medication periods, all smoked with no drug on week 1 (baseline or washout), began dose run-up on week 2, and tried to quit every day during week 3.

Setting: A university research center in the United States.

Participants: Forty-five adult smokers high in quit interest.

Measurements: Abstinence was verified daily each quit week by self-report of no smoking over the prior 24 hours and carbon monoxide (CO) < 5 parts per million.

Findings: Compared with placebo, bupropion did (F(1,44) = 6.98, P = 0.01), but modafinil did not (F(1,44) = 0.29, P = 0.60), increase the number of abstinent days. Also, bupropion (versus placebo) significantly increased the number of those able to maintain continuous abstinence on all 5 days throughout the quit week (11 versus four), Z = 2.11, P < 0.05, while modafinil did not (six).

Conclusions: Assessing days abstinent during 1 week of use of medication versus placebo in a cross-over design could be a useful early Phase 2 study design for discriminating between medications useful versus not useful in aiding smoking cessation.

Keywords: Bupropion; medication development; modafinil; pharmacotherapy; sensitivity; smoking cessation; specificity.

Figure 1

Mean (±SEM) days abstinent (left) and proportion able to achieve full continuous 5-day abstinence (right) during each quit week due to placebo, modafinil (100 mg b.i.d.), and bupropion (150 mg b.i.d.). All 45 subjects were included in each dependent measure in this within-subjects cross-over design. * p<.05 and ** p=.01 for significance between medication conditions.

Figure 2

Mean (±SEM) craving (QSU) and withdrawal (MNWS) scores due to placebo, modafinil, and bupropion. Data are only from days in which subjects were abstinent, to avoid confounding responses to medication with relief from continued smoking. * p<.05, ** p<.005, and *** p<.001 for significance from placebo. No other differences were significant.